A.I Wrote This Post… sorta

I'm utterly obsessed. In recent months, my excitement for ChatGPT, the chat-like generative AI making waves, has grown to border on obsession. It has become an indispensable ingredient in my daily endeavors—much like the way soy sauce makes its way into everything I cook.

Generative AI has been making headlines lately, and it's crucial to acknowledge that these tools are not flawless or genuinely intelligent. A recent incident involved a lawyer employing ChatGPT to draft a legal motion, only to discover that it was predominantly crafted by the AI itself. And there are plenty of exciting reasons to embrace these tools positively. On the Pivot Podcast, digital government champion Jen Pahlka shared a compelling tale of a civic designer utilizing ChatGPT to enhance a city's mailings and communication, making them more user-friendly.

Let me tell you about my…well, relationship with ChatGPT. 

Now that I’m finally willing to openly embrace my technical inclinations, let me begin with a personal example close to my heart—my smart home. Automating tasks in my living space has been a favorite hobby for a long time. The more my home can handle, the less mental effort I need to expend. However, complex automations require complex programming logic and syntax. 

Just last month, I gave ChatGPT with the following prompt: "I have a rain sensor called sensor.rain and a sprinkler controller named switch.zone1 and switch.zone2. Additionally, there's a weather forecast sensor labeled weather.bozeman. Can you craft an automation that waters the lawn at sunrise, except if it has rained more than 0.10" in the last 24 hours or rain is expected within the next two hours? In the case of expected rain, delay watering for two hours and recheck."

While sparing you the intricate code it generated, I can say it was perfect. Moreover, ChatGPT refined it in ways that would have eluded me. It even anticipated potential exceptions—such as the scenario where watering might be postponed until the next day—and accounted for them seamlessly.

What else can my new A.I. friend do you ask? 

A brief anecdote: I recently persuaded someone close to me to conquer one of North America's most challenging ski runs on our local mountain's closing day. TL;DR she has to have knee surgery soon. Don’t ask. Following her initial visit to the doctor, she contacted me in frustration, exclaiming, "I need an authorization before I can get an MRI, and they insist it must be sent by fax... what is an authorization and who is still using faxes?" All of this felt like a impediment to her getting back on her feet quickly. 

I turned to ChatGPT with a prompt: "Imagine an empowered, intelligent individual encountering the US healthcare system for the first time. What are some ways they could regain a sense of control?"

The suggestions ChatGPT provided closely aligned with my own recommendations, drawn from my extensive healthcare career. It even presented some ideas I hadn't considered—such as advising the person to contact their insurance company for an electronic authorization or inquiring about the most cost-effective location for an MRI, which, unsurprisingly, turned out to be the hospital rather than the orthopedic clinic. In the end, she secured an MRI on their terms and timeline.

My interactions with ChatGPT have gone way beyond practical matters. I've asked it about skiing, vegan meal ideas, troubleshooting 3D printing issues, and running routes. I even received a pretty entertaining standup routine,or 'tight five,' about our peculiar little mountain town. I've asked it for jokes tailored to a friend's birthday, incorporating specific details about the individual. Additionally, I fed it a list of my partner's preferred and disliked movies, and in return, it offered recommendations for new films. As you can tell, I'm entirely immersed in this obsession of mine.

Not only did ChatGPT pen this very post, but it also composed every other piece I've shared throughout the week.

Let me tell you about my process. During my morning runs, I contemplate various topics I wish to explore and share. During my cooldown walk, I employ text-to-speech and dictate a rough draft into a mobile app. Back home, I transfer the text to a Google Doc and commence the refining process—editing, refining, and adjusting as necessary.

This is when ChatGPT enters the picture. I seek its assistance for a summary and critique. The summary provides insight into whether my intended message is effectively conveyed or not, while the critique helps enhance flow and overall editing.

For instance, here's what ChatGPT had to say about the initial draft of this post: 

"Furthermore, the author mentions the AI's ability to provide a summary and critique but does not explicitly include them in the post. It would be beneficial to include an example summary and critique to demonstrate the AI's capabilities and showcase how it contributes to the writing process."

This is where ChatGPT gets really exciting for me. As a wordy writer, I often find it challenging to reimagine ways to convey my ideas once I've written a sentence. I rely on ChatGPT to provide edits, but before doing so, I kindly remind it of my nontraditional leadership background with a creative design focus. I ask it to use less jargon while ensuring the content remains respectful and avoids silliness or insulting language that may not resonate with the reader.

A genuine conversation ensues. Occasionally, I may request that ChatGPT reintroduce specific language that aligns with my style and voice (it once responded with, "Of course, that's your voice and style!"). Finally, I provide the prompt: "This will be posted on LinkedIn. Could you edit it for brevity, and if so, what suggestions do you have?"

And now, dear reader, you're looking at the result of this collaborative process.

I'll wrap by highlighting the true game changer for me. I've always loved writing, but, as I've confessed—and perhaps many of us can relate—I struggle to be my own editor. Collaborating with human editors hasn't always been smooth sailing either (my apologies to my Procedure Press friends!). At times, I find myself lost in the back-and-forth, and there are moments when I hold my darlings too dear to let go, as Ginsberg so aptly put it. 

However, with ChatGPT, I recognize that I'm working with a computer—an impartial entity. There's no ego involved, and I'm not obligated to accept its suggestions blindly. Much like my home automation endeavors and the healthcare advice it offers, I possess the discernment to judge when the AI is right or wrong.

Thanks to ChatGPT, the days of fretting over my abundant typos and spelling errors are a thing of the past.

I'm curious—anyone else obsessed with this thing?